POSITTRON: yet another proposal for positional head-tracking

This is for discussion and development of non-commercial open source VR/AR projects (e.g. Kickstarter applicable, etc). Contact MTBS admins at customerservice@mtbs3d.com if you are unsure if your efforts qualify.
German
Binocular Vision CONFIRMED!
Posts: 209
Joined: Tue Nov 23, 2010 5:18 pm

Re: POSITTRON: yet another proposal for positional head-trac

Post by German »

I don't know if you've seen it yet but check out the specs on the new PS4 Eye at the bottom:

Image
MSat
Golden Eyed Wiseman! (or woman!)
Posts: 1329
Joined: Fri Jun 08, 2012 8:18 pm

Re: POSITTRON: yet another proposal for positional head-trac

Post by MSat »

"Dedicated connector"

Bye bye PC interface capability (unless someone makes an interface box)
German
Binocular Vision CONFIRMED!
Posts: 209
Joined: Tue Nov 23, 2010 5:18 pm

Re: POSITTRON: yet another proposal for positional head-trac

Post by German »

MSat wrote:"Dedicated connector"

Bye bye PC interface capability (unless someone makes an interface box)
Or it's just USB 3 with a proprietary connector.
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

Yes that was one of the first things i checked after the reveal :-)

I guess it’s too early to tell if we’ll be able to use it on PC, but 640x400@120fps would be nice!
MSat
Golden Eyed Wiseman! (or woman!)
Posts: 1329
Joined: Fri Jun 08, 2012 8:18 pm

Re: POSITTRON: yet another proposal for positional head-trac

Post by MSat »

German wrote:
MSat wrote:"Dedicated connector"

Bye bye PC interface capability (unless someone makes an interface box)
Or it's just USB 3 with a proprietary connector.
I'm assuming the console will have actual USB3 ports. If so, then why would they make the connector for the camera proprietary?
German
Binocular Vision CONFIRMED!
Posts: 209
Joined: Tue Nov 23, 2010 5:18 pm

Re: POSITTRON: yet another proposal for positional head-trac

Post by German »

MSat wrote: I'm assuming the console will have actual USB3 ports. If so, then why would they make the connector for the camera proprietary?
It's basically AMD PC hardware and they need to keep costs as low as possible. The beginning of the new console cycle is the most expensive for console manufacturers. Coming up with a proprietary, high speed bus just for the camera would be expensive on both the development and manufacturing fronts. They already made that mistake with the PS3, it cost them a lot of money and they aren't likely to do it again.

If they want the camera plugged in all the time, which seems to be the way both Microsoft and Sony are going as far as motion controls, they wouldn't want consumers to use the general USB 3 ports for it. Having a dedicated connector that is different eliminates setup confusion as well as drawing a line that previous cameras don't work with the system. This is important when you have some accessories that are compatible(Move controller). The Kinect has a dedicated connector on newer Xboxes because it draws more power. You have to use an adapter and external power if you want to connect it to an older Xbox via USB. It's still USB though.
MSat
Golden Eyed Wiseman! (or woman!)
Posts: 1329
Joined: Fri Jun 08, 2012 8:18 pm

Re: POSITTRON: yet another proposal for positional head-trac

Post by MSat »

You may very well be right. Only time will tell.
User avatar
Namielus
Certif-Eyable!
Posts: 957
Joined: Thu Aug 02, 2012 8:49 am
Location: Norway
Contact:

Re: POSITTRON: yet another proposal for positional head-trac

Post by Namielus »

PatimPatam, how are you doing on this project? Is there anything you need?
Riftoholic

My precious 6 month project the Oculus Virtual Lounge:
Image
If you help me in any way I will be forever grateful.
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

Thanks for the interest Namielus, the truth is had a little break from this project during the last month. I’ve been doing some things that I neglected a bit before, like social life, playing the guitar and wedding preparations :-) Also started working on another VR related project which I cannot talk about for now..

But don’t worry i’m hoping i can get back to it soon, hopefully before i get my brand new dev kit in a few weeks (number 46XX!). First thing I want to do is test some new LEDs I just received by post, which hopefully will help improve the detection distance quite a bit!
User avatar
nateight
Sharp Eyed Eagle!
Posts: 404
Joined: Wed Feb 27, 2013 10:33 pm
Location: Youngstown, OH

Re: POSITTRON: yet another proposal for positional head-trac

Post by nateight »

Oculus, hire this man! (Or at least steal his tech. :lol:) PatimPatam, it's apparently true that no individual part of your design is really groundbreaking, but you've put it all together into a jaw-dropping package that would seem to require little additional R&D to be entirely at home in the commercial Rift. With vestibular organs so sensitive to the slightest discrepancies, I really don't think you can build 100% of the required components for proper positional tracking into an HMD (or at least one anyone but DARPA can afford). That leaves two options: 1) Building two cameras into the Rift and relying on stereoscopic cues for tracking (I'd love for this to be the winner, because it would accidentally make the Rift a combination virtual reality and mediated reality headset, but I just don't think existing visual odometry algorithms can possibly cover enough real world cases for this to be reliable in a consumer product - prove me wrong!) 2) Markers and one or more cameras. 2 is more likely to win because the existing algorithms are more numerous and more mature; I don't see users placing markers around their rooms to accommodate cameras in the Rift itself (though I'd still love to get AR for free), and with a good marker rig you only need one camera, so that's the ultimate resolution I'm betting on. Maybe Oculus has or is developing a pet solution to this problem that will blow us all away, but I'm not terribly hopeful - my prediction is you've predicted the tracking system of the commercial Rift. Bravo! :D

Now for the bad news - I don't see this selling more than maybe a few hundred units if it's an aftermarket device. Plenty of devs would buy it for their dev Rifts if they knew for sure they were getting something that would perfectly mimic what would eventually be built into the production model, but no amount of explaining why a consumer would want this is likely to make it catch on (short of half the people who use the Rift feeling ill from a lack of positional tracking, but if that happens we've all got bigger problems). Thing is, if we assume a cheap PS Eye-style camera is going to have to be sold alongside the Rift, building an array of LEDs into the housing is virtually free both from a parts standpoint and from an electronic standpoint. Integrate something like this into the design of the commercial Rift and I think you can fuse enough sensor data to produce head tracking that's at least adequate and probably much better.

Things I would continue to iterate? I don't have your CS chops to prove it, but I'm not certain you need all those LEDs (or more than a single color). If you assume this is going to be strictly for seated usage (a huge majority of Rift usage will be seated, IMO), and assume the swivel chair solution to avatar rotation will ultimately be proven naive, suddenly you may not need those troublesome LEDs in the back. Moreover, even if you have a system with which occlusion is possible, I've long wondered why motion capture systems and the like didn't introduce algorithms that simply "faked" the hidden data, made good-enough guesses based on last-known locations and vectors tied to a large sample of those motions while fully tracked, then picked up the markers when they reappeared and smoothly interpolated back to the sensed position - instead, the historical solution has essentially seen computers saying, "Where'd that one ball go?! I don't see that one ball OH GOD YOUR DATA IS NOW USELESS *CRASH*" Have the algorithms improved? It's 2013 here, guys. :lol: Something else to explore is using el wire to present lit lines to track instead of points - it would have the added benefit of looking even cooler, too. 8-)

All in all, thanks - as far as I'm concerned positional head tracking for seated gameplay in the Rift 1.0 is a solved problem, and considering the number of unsolved problems around here, that's a godsend. One wonders what we'll do for positional tracking while we wait for the Rift 1.0, of course, but I'm still hoping to get something halfway-usable out of two LED beacons and a 9-volt taped to a dev Rift and either Wiimotes or PS Eyes for tracking. Let me know if you intend to adapt your code to something like that because it's a problem on my list of things to tackle while I wait for #~8200 to be up - if I hit any major roadblocks when I get around to it myself you'll definitely be the first person I talk to. :D

EDIT: Clarified a word. Tracker != marker. Correct terminology is important and/or I'm dumb.
Shameless plug of the day - Read my witty comments on Reddit, in which I argue with the ignorant, over things that don't matter, for reasons I never fully understood!
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

Thanks a lot for the compliments nateight! I have to say that even though i couldn't participate on the discussion at the time i did see your youtube video on seated play controls, and i do share most of you ideas (even though i would implement it slightly differently).

For other people to see, here are the links in case you missed it!
Using feet for movement in a VR world
http://www.youtube.com/watch?v=EvtFn73LgDg

I know it is going against what many of the VR gurus around these forums think, but personally i believe that the near future of VR is on seated play (non-swiveling chairs please) and foot controls similar to your prototype.

So in relation to all this i guess you could ask... why did you bother with the back section of the PosiTTron then? The reason is that apart from the restriction of being seated itself i don't want the player to feel restricted in any way. Even while seated on a fixed chair you can easily turn your torso about 60 degrees, plus your head about 90 degrees: that makes 150 degrees either left or right.. a case were the back section would certainly be needed. Think walking on a lonely corridor on Doom 3 and suddenly hearing something creeping right behind you, i think the immediate instinct would be to turn your head/torso before turning your whole body.

And yeah my intention was never to sell this as a separate product really... i still want to try it myself when i get the oculus dev kit though, and i plan to share how to build the hardware + free software so people can try themselves if they want to. A simpler to build (but uglier) DIY solution could be using a couple of Move-type glowing balls, and attaching them Mickey Mouse style!
User avatar
Nick3DvB
Binocular Vision CONFIRMED!
Posts: 311
Joined: Wed Oct 06, 2010 10:51 am
Location: UK

Re: POSITTRON: yet another proposal for positional head-trac

Post by Nick3DvB »

I’m a bit late but I also want to say well done on this project! 8-) I’ve not been around for a while and sadly haven’t had the time to test any tracking ideas yet myself but it’s really great to see the progress you have made since we all first started looking into PTAM etc. I think the freedom your system gives is really imported even for seated play, I look forward to trying it out myself some day, we will have to have a good look at the SDK to see the best way to interface everything but hopefully it won’t be so hard. Really I think the only thing you got wrong is the name, (I like Ladytron to - Destroy is a great track) but looking at the shape you really should have called it the “ Kryten:lol:
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

Hahah, good geeky observation! Right up there with Mr Data's positronic brain!

Hmm I might consider changing the name then (but then again maybe not) :-P
User avatar
Nick3DvB
Binocular Vision CONFIRMED!
Posts: 311
Joined: Wed Oct 06, 2010 10:51 am
Location: UK

Re: POSITTRON: yet another proposal for positional head-trac

Post by Nick3DvB »

Fair enough... :lol:

Did you hear about the RasPi camera, it should be out very soon and can do 720p@60fps or VGA@90fps:

http://www.tweaktown.com/news/29201/mor ... index.html

Not sure that the Pi CPU has the power to run OpenCV locally though?
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

Hey thanks for the idea Nick, that is something well worth trying i think!

It is possible to run OpenCV on a RasPi, found a few projects around.. My guess is the CPU alone may not have enough power to run a smooth tracking, but i bet the GPU does (using CUDA). If i could make it work that would completely off-load the main PC, and open some new interesting possibilities! Apart from cost, the only problem could be the extra latency due to the added communication between the RasPi and the PC..

I could even test it using a RasPi and the same PS Eye, and if it works move to the RasPI camera you mentioned which looks like a nice improvement over Sony's one (increased resolution could mean extra precision and detection distance!).
User avatar
Nick3DvB
Binocular Vision CONFIRMED!
Posts: 311
Joined: Wed Oct 06, 2010 10:51 am
Location: UK

Re: POSITTRON: yet another proposal for positional head-trac

Post by Nick3DvB »

Sadly it doesn't look like you can use the GPU for much apart from video at the moment. :cry:

http://www.raspberrypi.org/phpBB3/viewt ... =24&t=3330
http://www.raspberrypi.org/phpBB3//view ... 68&t=12791
http://www.raspberrypi.org/phpBB3/viewtopic.php?p=14633

It would be fun to try it, but if we have to do the processing on the PC anyway we would not have enough output bandwidth for RAW video, and H264 compression would probably add too much latency? But I think we should still look into porting some of the OpenCV code to CUDA / OpenCL on the PC.

One other idea for you... I thought maybe we could use a bright Ultra-Violet lamp to illuminate the player (and maybe a UV filter on the camera?) and then UV stickers (or paint dots) on the HMD instead of LEDs? It would be easier to experiment with pattern changes, and maybe even a bit cheaper / easier to build, I will have to try it out some time. 8-)
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

Yep, looks like you're right. It's unfortunate that Broadcom haven't implemented an OpenCL driver for the VideoCore (and doesn't look like they will). To be honest that makes the RasPi much less appealing.. even if it was possible to send the video feed to a PC without much latency i think paying about $100 (raspi + cam board + case + power supply) for what would be basically a plain camera is a little bit too much.

I have thought about porting to CUDA / OpenCL on the PC but the problem is that apparently using CUDA while playing games does not bear good results.. so unless you buy an extra graphics card just to take care of the tracking.. (could be an option actually).

About the UV light idea, i'm not sure if it would be possible to discriminate between 3 or 4 different types of marker.. if i had more time in my hands sounds like it could be fun to test. Also not sure if it would be a good marketing strategy, forcing users to apply sunscreen before every VR session :-D
User avatar
Nick3DvB
Binocular Vision CONFIRMED!
Posts: 311
Joined: Wed Oct 06, 2010 10:51 am
Location: UK

Re: POSITTRON: yet another proposal for positional head-trac

Post by Nick3DvB »

I hope something happens about getting more access to the GPU in the future, if they sell millions broadcom might come around, or maybe just one genius with enough time in his hands. Even without the GPU it might still be a good IP camera for large scale multi-camera environments, using a power-over-ethernet adapter etc. Personally I think if we can get a ~30% speed-up using OpenCL on the PC side it’s worth a dedicated card, many have SLI systems now or use a old spare card for Nvidia PhysX, I know I would trade either for fast transnational tracking in a heart-beat!

UV lamps are so cheap now this could be an option for home use, but I was thinking more for clubs / laser-tag arenas that already have UV lighting installed. I would really like to play with UV when I get time, it should have excellent emissive uniformity and is very flexible, lots of different colors, different sized dots, shapes, lines etc. I do worry about the relative brightness in sunlight though, but we don’t get much of that here in the UK anyway! It’s ok for you in sunny Spain but I think many people here would be happy to play in VR and get a nice sun-tan at the same time!! :idea: 8-) :lol:
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

Hello there, quick update!

As many of you have seen, it has already been shown by tbowren that the Hydra solution is probably a more practical way to quickly hack-in positional tracking on the Rift for now:
http://www.mtbs3d.com/phpBB/viewtopic.php?f=140&t=17021
However I still want to test properly both approaches in order to evaluate the differences in accuracy, latency, range, etc.

Also is yet to be seen if 2 Hydra systems (4 controllers) can be used at once:
http://www.mtbs3d.com/phpBB/viewtopic.php?f=140&t=17035
So maybe the PosiTTron could be a good way to enable positional head-tracking and using 2 hydras for hand-tracking at the same time. Looks like adoral84 is already working on supporting 2 hands and head-tracking in his Half-Life 2 mod:
http://www.mtbs3d.com/phpBB/viewtopic.php?f=142&t=16174

Apart from that, as I mentioned before and some people have suggested (like EdZ or zalo for instance), the fact that we can take into account the orientation info from the Oculus IMU could simplify a lot the design of the PosiTTron and I’m thinking about going this route (in order to make it as easy as possible to DIY). I’m pretty confident we can go down to only 8 LEDs (one on each truncated vertex) and just 1 or 2 types of marker.. but that’s quite a radical change and I will need to reprogram a few important parts of the tracking software.

Just received my Razer Hydra yesterday and still waiting for my Rift. Unfortunately I don’t have much free time lately so I wouldn’t be surprised if someone else beats me to it!
User avatar
nateight
Sharp Eyed Eagle!
Posts: 404
Joined: Wed Feb 27, 2013 10:33 pm
Location: Youngstown, OH

Re: POSITTRON: yet another proposal for positional head-trac

Post by nateight »

Image

"Crystal Cove", they call it. There's something so familiar here, but I just can't put my finger on it... :lol:
Shameless plug of the day - Read my witty comments on Reddit, in which I argue with the ignorant, over things that don't matter, for reasons I never fully understood!
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

Yes well, it looks like it wasn't such a crazy idea after all. :-P

Kind of surprised they went for more LEDs instead of less, but i guess that's the price you have to pay for using IR LEDs instead of colored ones (and you still want to have some redundancy).

Would like to hear more details though, like how good is the robustness against occlusion (from your own hands/arms), or maximum tracking distance (could it be used in a living room for instance, putting the camera next to the TV and sitting on a sofa around 3 meters away from it?), or how much the tracking algorithm is going to tax the CPU/GPU (or will it be done by HW?).

Also what are the actual prototype camera specs?
- Resolution?
- FOV?
- FPS?

Finally, as i mentioned on the Oculus forum, having an external camera opens up some new interesting possibilities, like cheap controller/glove/hand positional tracking!
User avatar
brantlew
Petrif-Eyed
Posts: 2221
Joined: Sat Sep 17, 2011 9:23 pm
Location: Menlo Park, CA

Re: POSITTRON: yet another proposal for positional head-trac

Post by brantlew »

I would characterize it as being very robust against occlusion. The sheer number of LEDs helps in that regard. I can't make specific numerical comments but it works well in a variety of environments and lighting conditions. As long as you don't live in a hall of mirrors it should be perform well under most circumstances where you would be using a PC.
Zoide
Binocular Vision CONFIRMED!
Posts: 345
Joined: Fri Nov 18, 2011 1:32 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by Zoide »

brantlew wrote:I would characterize it as being very robust against occlusion. The sheer number of LEDs helps in that regard. I can't make specific numerical comments but it works well in a variety of environments and lighting conditions. As long as you don't live in a hall of mirrors it should be perform well under most circumstances where you would be using a PC.
What about a situation where you may turn around 180 degrees and face the other way? Would you need a second camera on the back?

Thanks
User avatar
brantlew
Petrif-Eyed
Posts: 2221
Joined: Sat Sep 17, 2011 9:23 pm
Location: Menlo Park, CA

Re: POSITTRON: yet another proposal for positional head-trac

Post by brantlew »

The device presented at CES is a single camera system and loses tracking if you turn 180 (reverting back to DK1-style head modeling). There are several instance in videos online where you can see tracking loss. In theory additional cameras could be used to compensate for wider angles and ranges.

That is all.
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

brantlew wrote:I would characterize it as being very robust against occlusion. The sheer number of LEDs helps in that regard. I can't make specific numerical comments but it works well in a variety of environments and lighting conditions.
Thanks a lot for the insight brantlew, that's very good news to hear; especially since motion tracked controllers are obviously a perfect fit for VR.
brantlew wrote:In theory additional cameras could be used to compensate for wider angles and ranges.
OR in theory they could make a fancier HMD design where LEDs would be visible even from the back (you don't need as many as in front since you don't have to deal with hands occlusion). Many possible options there.
michal
One Eyed Hopeful
Posts: 30
Joined: Mon Sep 03, 2012 5:46 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by michal »

I would assume that in most cases the user will be seated facing the computer screen (or just the camera).

In such a situation, unless you are possessed by a demon or are a contortionist, twisting your head 180 degrees or more would imply that you wouldn't have much flexibility left, thus reverting to normal neck modeling should feel completely realistic.

If you have more freedom of movement though, this falls apart.

I can't wait for the commercial Rift! Very excited.
lossofmercy
One Eyed Hopeful
Posts: 45
Joined: Thu Oct 18, 2012 2:25 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by lossofmercy »

Oculus could potentially sell the camera as it's own SKU and have the base sync with additional cameras wirelessly (given that it can figure out their position) to provide 360 coverage if people are interested in that. As long as it's easy to setup and mostly automatic, I think this is a really simple upgrade system for Oculus as it takes nothing away from a base model while adding functionality for the more demanding users.
michal
One Eyed Hopeful
Posts: 30
Joined: Mon Sep 03, 2012 5:46 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by michal »

lossofmercy wrote:Like I said in the developer forum, Oculus could potentially sell the camera as it's own SKU and have the base sync with additional cameras wirelessly (given that it can figure out their position) to provide 360 coverage if people are interested in that. I think this is a really simple upgrade system for Oculus as it takes nothing away from a base model.
It would require a considerable additional investment of resources in the software. Also wireless cameras would add latency.

It wouldn't surprise me if this is supported eventually, likely with wired cameras, but not in time for the commercial Rift. The proportion of people who would benefit is tiny. Probably the same people who have rooms dedicated to VR and intend on getting walking platforms, etc.

Of course, just like everyone else, I'm speculating (based on common sense).
lossofmercy
One Eyed Hopeful
Posts: 45
Joined: Thu Oct 18, 2012 2:25 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by lossofmercy »

Do wireless cameras add that much more latency? I hope not, it would make my life a lot simpler.
The proportion of people who would benefit is tiny. Probably the same people who have rooms dedicated to VR and intend on getting walking platforms, etc.
I disagree, I think a lot of people would really want it. Think about all the people that bought the Omni or think about developers like Cloudhead games who are trying to make VR games and really desire 360 position tracking (or really any game that has you walking around). They would need only one additional camera to make this possible. The most complicated thing that they would have to do, in addition to what they have already done, is a software that figures out the position of the camera in relation to the rift. I think they are already 80% there.
michal
One Eyed Hopeful
Posts: 30
Joined: Mon Sep 03, 2012 5:46 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by michal »

lossofmercy wrote:Do wireless cameras add that much more latency? I hope not, it would make my life a lot simpler.
The proportion of people who would benefit is tiny. Probably the same people who have rooms dedicated to VR and intend on getting walking platforms, etc.
I disagree, I think a lot of people would really want it. Think about all the people that bought the Omni or think about developers like Cloudhead games who are trying to make VR games and really desire 360 position tracking. They would need only one additional camera to make this possible. The most complicated thing that they would have to do, in addition to what they have already done, is a software that figures out the position of the camera in relation to the rift. I think they are already 80% there.
I think it will happen, but not at the outset. I think Oculus is initially targeting normal people who will sit in a stationary chair, and the number of people who have or will buy the Omni or something similar is going to be MAYBE 1% of those who buy the Rift. I think I'm being generous.

I would bet that many more people who buy the Rift will also buy (or have already bought) a steering wheel or joystick peripheral.

I hadn't heard anything about Cloudhead games desiring a 360 solution. Do you have a link? I can see how spelunking could benefit from that, though I would be somewhat afraid to allow myself that much freedom of movement in my room :P
lossofmercy
One Eyed Hopeful
Posts: 45
Joined: Thu Oct 18, 2012 2:25 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by lossofmercy »

http://www.reddit.com/r/oculus/comments ... pe/cejryfm
CloudHeadGames wrote:This would seem to limit the potential of a number of game play styles, especially in the first person. Rotational translations with analog and mouse tend to be the worst offenders in making people feel sick. We would love to get rid of right stick analog and even mouse rotations in favor of a "stand up and turn around" solution. If we are limited by 180 turns, that's a bit of a disappointment to be honest.
This is more of an issue when you have motion control systems in play of course.
That's the thing, they can target the masses with the basic one camera setup. They look like they already have it mostly down that they could go for an easy upgrade option. All they are losing is development time to sync the data between two cameras. And if it's easy to setup, then people would buy it as well. Well... if it's wireless. :D
michal
One Eyed Hopeful
Posts: 30
Joined: Mon Sep 03, 2012 5:46 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by michal »

lossofmercy wrote:http://www.reddit.com/r/oculus/comments ... pe/cejryfm
CloudHeadGames wrote:This would seem to limit the potential of a number of game play styles, especially in the first person. Rotational translations with analog and mouse tend to be the worst offenders in making people feel sick. We would love to get rid of right stick analog and even mouse rotations in favor of a "stand up and turn around" solution. If we are limited by 180 turns, that's a bit of a disappointment to be honest.
This is more of an issue when you have motion control systems in play of course.
That's the thing, they can target the masses with the basic one camera setup. They look like they already have it mostly down, that they could go for an easy upgrade option. All they are losing is development time to sync the data between two cameras. Well... if it's wireless. :D

Thanks for that.

I think the bigger hurdle (rather than wireless cameras), would be a wireless HMD. If you spin around a few thousand degrees the cables are going to get tangled regardless of how you mount them. Wired cameras are a much easier problem to deal with since they'd be stationary.

I'm still certain most people won't make use of more than 180 degrees though. Just as most people don't make use of the Kinect or PS Move due to lack of space. Also the reason people abandoned music games due to peripherals and dance mats taking up too much space.

I find the steering wheel I have to be a pain to mount and unmount over and over again and storing the pedals is also annoying. The proportion of people who will benefit from more than 180 degrees is likely the same people who have racing seats for sim racing. ie too small to bother catering for in the first iteration of the commercial product. Or maybe there will be a Rift Pro for commercial/research purposes which has additional features?

These are awesome problems to be discussing as I didn't expect things to come so far so quickly :)
lossofmercy
One Eyed Hopeful
Posts: 45
Joined: Thu Oct 18, 2012 2:25 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by lossofmercy »

Yup, wireless Rift would be awesome. Not expecting to see that in the ~300 price range for a while though.
User avatar
Tbone
Two Eyed Hopeful
Posts: 64
Joined: Tue Jan 15, 2013 5:43 pm
Location: Washington, DC
Contact:

Re: POSITTRON: yet another proposal for positional head-trac

Post by Tbone »

CloudHeadGames wrote:This would seem to limit the potential of a number of game play styles, especially in the first person. Rotational translations with analog and mouse tend to be the worst offenders in making people feel sick. We would love to get rid of right stick analog and even mouse rotations in favor of a "stand up and turn around" solution. If we are limited by 180 turns, that's a bit of a disappointment to be honest.
This is more of an issue when you have motion control systems in play of course.
I have to agree that I'd like to be able to eliminate rotation with the right stick. FPS type games are definitely a tougher problem than cockpit/seated games. Even if I play seated, I'd like to be able to push my chair back and swivel around 360 degrees (which can spark a discussion about the wires as well). It's so easy to get lost in the virtual world, even if your intention is to just stay seated and play facing forward, chances are you'll eventually get turned around 180, so this does seem like an issue that will need addressing.
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

michal wrote:
lossofmercy wrote:
The proportion of people who would benefit is tiny. Probably the same people who have rooms dedicated to VR and intend on getting walking platforms, etc.
I disagree, I think a lot of people would really want it. Think about all the people that bought the Omni or think about developers like Cloudhead games who are trying to make VR games and really desire 360 position tracking. They would need only one additional camera to make this possible. The most complicated thing that they would have to do, in addition to what they have already done, is a software that figures out the position of the camera in relation to the rift. I think they are already 80% there.
I think it will happen, but not at the outset. I think Oculus is initially targeting normal people who will sit in a stationary chair, and the number of people who have or will buy the Omni or something similar is going to be MAYBE 1% of those who buy the Rift. I think I'm being generous.

I would bet that many more people who buy the Rift will also buy (or have already bought) a steering wheel or joystick peripheral.

I hadn't heard anything about Cloudhead games desiring a 360 solution. Do you have a link? I can see how spelunking could benefit from that, though I would be somewhat afraid to allow myself that much freedom of movement in my room :P
As i have mentioned before i'm personally not very interested in omni treadmills or physically turning in VR at the moment.. however i think being able to look behind you 180 degrees without breaking the positional tracking would be very useful even for games where you are sitting down (or standing up without turning your whole body):

https://twitter.com/oculus/status/314251208040792064
michal
One Eyed Hopeful
Posts: 30
Joined: Mon Sep 03, 2012 5:46 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by michal »

Tbone wrote:
CloudHeadGames wrote:This would seem to limit the potential of a number of game play styles, especially in the first person. Rotational translations with analog and mouse tend to be the worst offenders in making people feel sick. We would love to get rid of right stick analog and even mouse rotations in favor of a "stand up and turn around" solution. If we are limited by 180 turns, that's a bit of a disappointment to be honest.
This is more of an issue when you have motion control systems in play of course.
I have to agree that I'd like to be able to eliminate rotation with the right stick. FPS type games are definitely a tougher problem than cockpit/seated games. Even if I play seated, I'd like to be able to push my chair back and swivel around 360 degrees (which can spark a discussion about the wires as well). It's so easy to get lost in the virtual world, even if your intention is to just stay seated and play facing forward, chances are you'll eventually get turned around 180, so this does seem like an issue that will need addressing.

This is also why I wish my swivel chair had a lock. There's just no way to avoid swiveling if your chair allows it. I'm most excited about cockpit style games or virtual experiences within VR that don't rely as much on realism and complete freedom of movement.
michal
One Eyed Hopeful
Posts: 30
Joined: Mon Sep 03, 2012 5:46 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by michal »

PatimPatam wrote: As i have mentioned before i'm personally not very interested in omni treadmills or physically turning in VR at the moment.. however i think being able to look behind you 180 degrees without breaking the positional tracking would be very useful even for games where you are sitting down (or standing up without turning your whole body):

https://twitter.com/oculus/status/314251208040792064
Given the neck model, I'm sure looking behind you will still feel completely natural. There is little need for positional tracking when your neck is craned 180 degrees backwards. The other sensors should be able to maintain the immersion perfectly.
User avatar
PatimPatam
Binocular Vision CONFIRMED!
Posts: 214
Joined: Thu Jun 28, 2012 1:31 pm
Location: Barcelona

Re: POSITTRON: yet another proposal for positional head-trac

Post by PatimPatam »

michal wrote:
PatimPatam wrote: As i have mentioned before i'm personally not very interested in omni treadmills or physically turning in VR at the moment.. however i think being able to look behind you 180 degrees without breaking the positional tracking would be very useful even for games where you are sitting down (or standing up without turning your whole body):

https://twitter.com/oculus/status/314251208040792064
Given the neck model, I'm sure looking behind you will still feel completely natural. There is little need for positional tracking when your neck is craned 180 degrees backwards. The other sensors should be able to maintain the immersion perfectly.
Sorry but I'm not so sure about that.. in the position shown in the picture you can still easily lean your body (trunk) to change your perspective, same thing if you're standing up and shooting things that come behind you for instance, nothing to do with the neck model.
michal
One Eyed Hopeful
Posts: 30
Joined: Mon Sep 03, 2012 5:46 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by michal »

PatimPatam wrote:
michal wrote:
PatimPatam wrote: As i have mentioned before i'm personally not very interested in omni treadmills or physically turning in VR at the moment.. however i think being able to look behind you 180 degrees without breaking the positional tracking would be very useful even for games where you are sitting down (or standing up without turning your whole body):

https://twitter.com/oculus/status/314251208040792064
Given the neck model, I'm sure looking behind you will still feel completely natural. There is little need for positional tracking when your neck is craned 180 degrees backwards. The other sensors should be able to maintain the immersion perfectly.
Sorry but I'm not so sure about that.. in the position shown in the picture you can still easily lean your body (trunk) to change your perspective, same thing if you're standing up and shooting things that come behind you for instance, nothing to do with the neck model.
The way I understand the neck model, you cannot lean without also tilting (unless you break your neck), thus the model can work out what is going on accurately. The only exception would be standing up or sitting down with your head craned back.

EDIT: Just to clarify, this is true only when your head is craned back at 180 degrees. Obviously you can lean without tilting otherwise.
lossofmercy
One Eyed Hopeful
Posts: 45
Joined: Thu Oct 18, 2012 2:25 am

Re: POSITTRON: yet another proposal for positional head-trac

Post by lossofmercy »

It would be nice to have actual position instead of predicted position using neck simulation, even if it is something you use when backing up.
Post Reply

Return to “VR/AR Research & Development”